Informing infant and young child feeding programming in humanitarian emergencies: An evidence map of reviews including low and middle income countries.

Around 200 million people were affected by conflict and natural disasters in 2015. Whereas those populations are at a particular high risk of death, optimal breastfeeding and complementary feeding practices could prevent almost 20% of deaths amongst children less than 5 years old. Yet, coverage of interventions for improving infant and young child feeding (IYCF) practices in emergencies is low, partly due to lack of evidence. Considering the paucity of data generated in emergencies to inform programming, we conducted an evidence map from reviews that included low- and middle-income countries and looked at several interventions: (a) social and behavioural change interpersonal and mass communication for promoting breastfeeding and adequate complementary feeding; (b) provision of donated complementary food; (c) home-based fortification with multiple micronutrient powder; (d) capacity building; (e) cash transfers; (f) agricultural or fresh food supply interventions; and (g) psychological support to caretakers. We looked for availability of evidence of these interventions to improve IYCF practices and nutritional status of infants and young children. We identified 1,376 records and included 28 reviews meeting the inclusion criteria. The highest number of reviews identified was for behavioural change interpersonal communication for promoting breastfeeding, whereas no review was identified for psychological support to caretakers. We conclude that any further research should focus on the mechanisms and delivery models through which effectiveness of interventions can be achieved and on the influence of contextual factors. Efforts should be renewed to generate evidence of effectiveness of IYCF interventions during humanitarian emergencies despite the challenges.

Public health information in crisis-affected populations: a review of methods and their use for advocacy and action.

Valid and timely information about various domains of public health underpins the effectiveness of humanitarian public health interventions in crises. However, obstacles including insecurity, insufficient resources and skills for data collection and analysis, and absence of validated methods combine to hamper the quantity and quality of public health information available to humanitarian responders. This paper, the second in a Series of four papers, reviews available methods to collect public health data pertaining to different domains of health and health services in crisis settings, including population size and composition, exposure to armed attacks, sexual and gender-based violence, food security and feeding practices, nutritional status, physical and mental health outcomes, public health service availability, coverage and effectiveness, and mortality. The paper also quantifies the availability of a minimal essential set of information in large armed conflict and natural disaster crises since 2010: we show that information was available and timely only in a small minority of cases. On the basis of this observation, we propose an agenda for methodological research and steps required to improve on the current use of available methods. This proposition includes setting up a dedicated interagency service for public health information and epidemiology in crises.

Using cross-sectional surveys to estimate the number of severely malnourished children needing to be enrolled in specific treatment programmes.

OBJECTIVE: When planning severe acute malnutrition (SAM) treatment services, estimates of the number of children requiring treatment are needed. Prevalence surveys, used with population estimates, can directly estimate the number of prevalent cases but not the number of subsequent incident cases. Health managers often use a prevalence-to-incidence conversion factor (J) derived from two African cohort studies to estimate incidence and add the expected number of incident cases to prevalent cases to estimate expected SAM caseload for a given period. The present study aimed to estimate J empirically in different contexts. DESIGN: Observational study, with J estimated by correlating expected numbers of children to be treated, based on prevalence surveys, population estimates and assumed coverage, with the observed numbers of SAM patients treated. SETTING: Survey and programme data from six African and Asian countries. SUBJECTS: Twenty-four data sets including prevalence surveys and programme admissions data for 5 months following the survey. RESULTS: A statistically significant relationship between the number of SAM cases admitted to SAM treatment services and the estimated burden of SAM from prevalence surveys was found. Estimate for the slope (intercept forced to be zero) was 2·17 (95 % CI 1·33, 3·79). Estimates for the prevalence-to-incidence conversion factor J varied from 2·81 to 11·21, assuming programme coverage of 100 % and 38 %, respectively. CONCLUSIONS: Estimation of expected caseload from prevalence may require revision of the currently used prevalence-to-incidence conversion factor J of 1·6. Appropriate values for J may vary between different locations.

Effect of ready-to-use foods for preventing child undernutrition in Niger: analysis of a prospective intervention study over 15 months of follow-up.

Strategies for preventing undernutrition comprise a range of interventions, including education, provision of complementary food and cash transfer. Here, we compared monthly distributions of two different lipid-based nutrient supplements (LNS), large-quantity LNS (LNS-LQ) and medium-quantity LNS (LNS-MQ) for 15 months on prevention of undernutrition among children 6 to 23 months. Both groups also received cash transfer for the first 5 months of the intervention. We conducted a prospective intervention study in Maradi, Niger, between August 2011 and October 2012. Six and 11 villages were randomly allocated to LNS-LQ/Cash and LNS-MQ/Cash, respectively. Children measuring 60-80 cm were enrolled in the respective groups and followed up monthly. Poisson regression was used to assess differences between interventions and adjust for baseline characteristics, intervention periods and child-feeding practices. The analysis included 2586 children (1081 in the LNS-LQ/Cash group and 1505 in the LNS-MQ/Cash group). This study suggests that provision of LNS-LQ (reference) or LNS-MQ had, overall, similar effect on incidence of severe acute malnutrition (RR = 0.97; 95% CI: 0.67-1.40; P = 0.88), moderate acute malnutrition (RR = 1.20; 95% CI: 0.97-1.48; P = 0.08), severe stunting (RR = 0.94; 95% CI: 0.70-1.26; P = 0.69), moderate stunting (RR = 0.95; 95% CI: 0.76-1.19; P = 0.67) and mortality (RR = 0.83; 95% CI: 0.41-1.65; P = 0.59). Compared with LNS-LQ, LNS-MQ showed a greater protective effect on moderate acute malnutrition among children with good dietary adequacy: RR = 0.72; 95% CI: 0.56-0.94; P = 0.01. These results highlight the need to design context-specific programmes. Provision of LNS-LQ might be more appropriate when food insecurity is high, while when food security is better, distribution of LNS-MQ might be more appropriate.

Using mid-upper arm circumference to end treatment of severe acute malnutrition leads to higher weight gains in the most malnourished children.

OBJECTIVE: The World Health Organization recommends discharging children admitted to nutrition programs treating severe acute malnutrition, with a low mid-upper arm circumference (MUAC <115 mm) when weight gain is >15%. When this recommendation is followed, the most severely malnourished children receive a shorter treatment compared to children that are less severely malnourished. This study assesses whether using MUAC >125 mm as discharge criteria eliminates this effect. METHODS AND FINDINGS: Data from 753 children cured from a Médecins Sans Frontières outpatient nutrition program in Gedaref, North Sudan were analyzed. MUAC >125 mm was used as discharge criteria. Length of stay and percent weight gain of children were compared in relation to nutritional status on admission. Children with low MUAC on admission had a longer duration of treatment (p = 0.000) and also a higher percent weight gain (p = 0.000) than children with higher MUAC. Similar results with weight-for-height z-scores categories were shown with both duration of treatment (p = 0.000) and percent weight gain (p = 0.000). CONCLUSION: This study shows that using MUAC as the discharge criteria eliminates the effect of shorter treatment in most severely malnourished children compared to least severely malnourished, as is observed with percent weight gain. The findings directly address the main concern that has been identified with the current WHO recommendation of using percent weight gain. MUAC could be used as discharge criteria, instead of percent weight gain, as having a longer duration of treatment and a higher percent weight gain for the most malnourished is highly desirable.

Assessment of the PROBIT approach for estimating the prevalence of global, moderate and severe acute malnutrition from population surveys.

OBJECTIVE: Prevalence of acute malnutrition is classically estimated by the proportion of children meeting a case definition in a representative population sample. In 1995 the WHO proposed the PROBIT method, based on converting parameters of a normally distributed variable to cumulative probability, as an alternative method requiring a smaller sample size. The present study compares classical and PROBIT methods for estimating the prevalence of global, moderate and severe acute malnutrition (GAM, MAM and SAM) defined by weight-for-height Z-score (WHZ) or mid-upper arm circumference (MUAC). DESIGN: Bias and precision of classical and PROBIT methods were compared by simulating a total of 1·26 million surveys generated from 560 nutrition surveys. SETTING: Data used for simulation were derived from nutritional surveys of children aged 6-59 months carried out in thirty-one countries around the world. SUBJECTS: Data of 459 036 children aged 6-59 months from representative samples were used to generate simulated populations. RESULTS: The PROBIT method provided an estimate of GAM, MAM and SAM using WHZ or MUAC proportional to the true prevalence with a small systematic overestimation. The PROBIT method was more precise than the classical method for estimating the prevalence for GAM, MAM and SAM by WHZ or MUAC for small sample sizes (i.e. n<150 for SAM and GAM; n<300 for MAM), but lost this advantage when sample sizes increased. CONCLUSIONS: The classical method is preferred for estimating acute malnutrition prevalence from large sample surveys. The PROBIT method may be useful in sentinel-site surveillance systems with small sample sizes.

Measuring anthropometric indicators through nutrition surveillance in humanitarian settings: options, issues, and ways forward.

The technical discourse on nutrition surveillance started decades ago, and the first technical guidelines were proposed in mid-1970s. In spite of this long history, little evidence and consensus exists on the best methods for conducting nutrition surveillance, and on the validity of data produced by these approaches. Multiple nutrition surveillance systems exist in humanitarian settings; however, the validity and usefulness of data produced by these systems are often questionable. In this paper, we outline and define five major methodological approaches to collecting child anthropometric data through surveillance: repeated surveys, community-based sentinel sites, mass screenings, admission data from feeding centers, and data from health clinics. We discuss outstanding methodological and practical challenges with direct implications for quality, validity, and interpretability of collected data and highlight comparative advantages and disadvantages of different methods. We also propose ways forward to building a better evidence base by documenting the strengths and limitations of different approaches, with the eventual goal of achieving consensus on the best ways to collect anthropometric data through surveillance.

An algorithm to assess methodological quality of nutrition and mortality cross-sectional surveys: development and application to surveys conducted in Darfur, Sudan.

BACKGROUND: Nutrition and mortality surveys are the main tools whereby evidence on the health status of populations affected by disasters and armed conflict is quantified and monitored over time. Several reviews have consistently revealed a lack of rigor in many surveys. We describe an algorithm for analyzing nutritional and mortality survey reports to identify a comprehensive range of errors that may result in sampling, response, or measurement biases and score quality. We apply the algorithm to surveys conducted in Darfur, Sudan. METHODS: We developed an algorithm based on internationally agreed upon methods and best practices. Penalties are attributed for a list of errors, and an overall score is built from the summation of penalties accrued by the survey as a whole. To test the algorithm reproducibility, it was independently applied by three raters on 30 randomly selected survey reports. The algorithm was further applied to more than 100 surveys conducted in Darfur, Sudan. RESULTS: The Intra Class Correlation coefficient was 0.79 for mortality surveys and 0.78 for nutrition surveys. The overall median quality score and range of about 100 surveys conducted in Darfur were 0.60 (0.12-0.93) and 0.675 (0.23-0.86) for mortality and nutrition surveys, respectively. They varied between the organizations conducting the surveys, with no major trend over time. CONCLUSION: Our study suggests that it is possible to systematically assess quality of surveys and reveals considerable problems with the quality of nutritional and particularly mortality surveys conducted in the Darfur crisis.

Trends in malnutrition and mortality in Darfur, Sudan, between 2004 and 2008: a meta-analysis of publicly available surveys.

BACKGROUND: The humanitarian response to the crisis in Darfur is the largest humanitarian operation in the world. To investigate the evolution of the conditions of the affected population, we analysed trends in malnutrition and mortality, the most widely accepted indicators for assessing the degree of severity of a crisis. METHODS: We did a meta-analysis of 164 publicly available surveys taking into account changes in the contextual situation and humanitarian aid; type of population [residents and internally displaced persons (IDPs)]; and seasonal variations. Data on global acute malnutrition (GAM), severe acute malnutrition (SAM), crude death rate (CDR) and under-five death rate (U5DR) were analysed using a random effect model. RESULTS: GAM and SAM decreased by 16% and 28%, respectively, in 2004-05, whereas CDR dropped by 44-75% per year depending on state and type of population and U5DR decreased by an overall 50% yearly. Both security and the humanitarian contexts became increasingly complex after 2005, but levels of malnutrition stabilized in North and South Darfur. In West Darfur, GAM remained stable but SAM tended to increase for IDPs, although mortality rates remained constant. Mortality increased slightly for residents in South Darfur after 2005, even though nutritional status was stable. GAM, SAM, CDR and U5DR fluctuated markedly with seasons. CONCLUSION: A meta-analysis of myriads of surveys permitted us to draw an overall picture of the situation in Darfur and to identify some of its influencing factors. The large humanitarian operation, which gained momentum through 2004-05, was able to contain the crisis despite huge difficulties, but did not compensate for seasonal variations. The situation has remained fragile with some negative patterns tending to emerge. It is crucial that the humanitarian situation continues to be closely monitored.

A review of methodology and analysis of nutrition and mortality surveys conducted in humanitarian emergencies from October 1993 to April 2004.

BACKGROUND: Malnutrition prevalence and mortality rates are increasingly used as essential indicators to assess the severity of a crisis, to follow trends, and to guide decision-making, including allocation of funds. Although consensus has slowly developed on the methodology to accurately measure these indicators, errors in the application of the survey methodology and analysis have persisted. The aim of this study was to identify common methodological weaknesses in nutrition and mortality surveys and to provide practical recommendations for improvement. METHODS: Nutrition (N = 368) and crude mortality rate (CMR; N = 158) surveys conducted by 33 non-governmental organisations and United Nations agencies in 17 countries from October 1993 to April 2004 were analysed for sampling validity, precision, quality of measurement and calculation according to several criteria. RESULTS: One hundred and thirty (35.3%) nutrition surveys and 5 (3.2%) CMR surveys met the criteria for quality. Quality of surveys varied significantly depending on the agency. The proportion of nutrition surveys that met criteria for quality rose significantly from 1993 to 2004; there was no improvement for mortality surveys during this period. CONCLUSION: Significant errors and imprecision in the methodology and reporting of nutrition and mortality surveys were identified. While there was an improvement in the quality of nutrition surveys over the years, the quality of mortality surveys remained poor. Recent initiatives aimed at standardising nutrition and mortality survey quality should be strengthened. There are still a number of methodological issues in nutrition and mortality surveys in humanitarian emergencies that need further study.